模型与评测

Anthropic takes steps to block Claude from reverse-engineering video games

ATLAS-Bench: An Agentic Tool-Use Benchmark for Local Area Search

arXiv:2610.39438v1 Announce Type: new Abstract: Reliable local search is essential for LLM agents to support everyday decisions and longer plans by finding places that meet users' needs at given times and locations. Yet existing benchmarks provide limited coverage of whether sear

AgentDecept: Benchmarking Deceptive Behavior in Tool-Using LLM Agents under Pressure

arXiv:2610.39377v1 Announce Type: new Abstract: LLM agents increasingly carry out multi-turn tasks across everyday applications, yet users often see only the agent's report rather than the observations, API calls, and state changes behind it. Existing deception benchmarks have st

Making it easier for websites to express AI training preferences

OpenAI will honor a shared robots.txt signal to help website operators communicate their AI training preferences to participating providers.

An Introduction to Large Language Models for Scientists and Professionals

arXiv:2610.37684v1 Announce Type: cross Abstract: Large Language Models (LLMs) have rapidly become important tools for science, education, medicine, and professional work. Despite their apparent simplicity from the user's perspective, effective and responsible use requires an und

Answer-Only Disclosure: Private-Context LLM Assistance over Web APIs with a Leakage Ledger

arXiv:2610.38126v1 Announce Type: cross Abstract: Web assistants increasingly answer questions about a user's private documents by pairing a small on-device model with a cloud LLM behind a Web API. In collaborative decoding, the document stays on the client, and the cloud receive

Beyond LLM Wrappers: Context Engineering and the Three-Engine Architecture for Reliable Financial AI

arXiv:2610.39230v1 Announce Type: new Abstract: Financial AI systems must support decision-bearing questions involving payments, collections, cash flow, and compliance while ensuring numerical accuracy, traceability, and auditability. Large language models (LLMs) enable natural l

Beyond fabrication: Misattributed numbers in papers written by AI research agents

arXiv:2610.38691v1 Announce Type: new Abstract: Verifying numerical claims in research written by AI agents is an important but challenging problem. Recent benchmarks show that many agent-written papers fail integrity checks, but such verdicts, often produced by a single LLM judg

BiShield-TEE: On the (In-)Security of Unilateral Weight Obfuscation in On-Device TEE-Shielded LLM Partition

arXiv:2610.21565v3 Announce Type: replace-cross Abstract: On-device deployment of large language models (LLMs) is increasingly adopted due to latency and privacy requirements, but exposes proprietary models to reverse engineering and intellectual property theft. Trusted Execution

Can LLMs Faithfully Solve Ontology Engineering Tasks? A Mechanistic Interpretability Study

arXiv:2610.38326v1 Announce Type: new Abstract: Ontological knowledge underpins large language model (LLM) performance on knowledge-intensive tasks. However, existing evaluations of LLMs' ontological understanding largely rely on behavioral correctness, leaving the underlying mec

Comparing Human Role-Players and LLM-Simulated Clients in Online Counselling Training: An Analysis of Counselling Patterns

arXiv:2610.40563v1 Announce Type: new Abstract: This study investigates the capabilities of Large Language Models to simulate counselling clients in educational role-plays in comparison to human role-players. Initially, we recorded role-playing sessions, where novice counsellors

Do Personalized Skills Help Coding Agents? An Empirical Study of Developer Interaction Traces

arXiv:2608.10319v3 Announce Type: replace-cross Abstract: As human developers increasingly collaborate with large language model (LLM)-powered coding agents, adapting to individual developer preferences, potentially via agent skills, may offer way to reduce repeated guidance and

Epistemic Warrant for LLM Recommendations: Characterizing the Basis for Reliance When Ground Truth Is Unavailable

arXiv:2609.04127v2 Announce Type: replace Abstract: Large language models are increasingly used to support organizational decisions. Despite their rapid and widespread adoption, users often lack a principled basis for assessing whether to rely on a specific recommendation. We int

HEMO: Harness Self-Evolution for Multi-Objective Combinatorial Optimization

arXiv:2610.40023v1 Announce Type: cross Abstract: Automatic operator design (AOD) has increasingly adopted large language models (LLMs) to generate problem-specific search and variation operators for multi-objective combinatorial optimization (MOCO). Yet most methods evolve the o

Interpretable Feature Discovery from Unstructured Data Using LLMs and Statistical Choice Models

arXiv:2610.38082v1 Announce Type: cross Abstract: Feature extraction from unstructured data typically relies on opaque embeddings unsuited to interpretable decision-making. Interpretable features, on the other hand, typically require costly, and often infeasible, hand-engineering

LLM-Assisted Optimization for Simulation-Based Redesign

arXiv:2610.39275v1 Announce Type: new Abstract: Engineering design frequently involves redesigning a prior solution to adapt to changes in context, such as material prices or system requirements. Redesign often leverages simulators too complex and/or computationally intensive for

Masked Erosion: How Acceptance Gates Lose Redundant Safeguards in Self-Improving Agents

arXiv:2610.38392v1 Announce Type: new Abstract: Self-improving LLM agents edit their own harness, the prompts and rules around a fixed model, and keep an edit only if an acceptance gate scores it higher on a batch of tasks. Such a gate sees what the harness does under the current

Poisoning Agentic Alpha: Adversarial Vulnerabilities Across Roles and Architectures in Multi-Agent Trading Systems

arXiv:2608.24069v2 Announce Type: replace Abstract: LLM-based multi-agent trading systems are moving from research prototypes to live deployments that control real assets. They ingest news and social media from the open web and pass it between agents, so a single corrupted signal

SCOPE: Selective Critical-Step On-Policy Self-Distillation for Agentic Reinforcement Learning

arXiv:2610.38391v1 Announce Type: new Abstract: Outcome-based reinforcement learning provides coarse supervision for long-horizon LLM agents: failed trajectories may contain correct steps, while successful trajectories can contain errors. Trajectory-level skills can also shift te

The Athlete as a Scientist of One: Agentic Bayesian Experimental Design for Individualized Physiology from Consumer Wearables

arXiv:2610.40370v1 Announce Type: cross Abstract: Large language model (LLM) agents for personal health read wearable data and recommend, but do not learn about the individual in a way that can be audited or falsified. We present EpiCoach, an agentic system that treats one athlet

The Real Cost of Context: Predicting Compute Time and Energy Consumption in LLM Prefill and Decode Phases

arXiv:2610.39621v1 Announce Type: new Abstract: The integration of Large Language Models (LLMs) into agentic architectures has substantially increased token consumption, raising concerns about their energy footprint. Yet, estimating the energy consumption of proprietary API-based

Towards AI-Assisted Clinical Trial Matching: Practical Considerations, Multicenter Evaluation, and Real-World Deployment

arXiv:2609.01202v2 Announce Type: replace-cross Abstract: Clinical trials are essential for evidence-based medicine and drug development, but many fail because of insufficient patient enrollment. Although there is growing interest in using AI to support patient recruitment, exist

TrussTutor-Bench: Benchmarking large language models on solution accuracy, sycophancy, and error detection in statics

arXiv:2610.38917v1 Announce Type: cross Abstract: Large language models (LLMs) have rapidly become a default source of help for students in science and engineering courses. In foundational engineering mechanics courses such as statics, students routinely use these tools to check

Well-Specified and Well-Tested? Assessing SWE-bench-Style Datasets with a Self-Evolving Agent

arXiv:2610.39417v1 Announce Type: cross Abstract: Large language models (LLMs) have shown strong performance on software engineering (SE) tasks, driving the need for high-quality datasets for reliable training and evaluation. SWE-bench-style datasets (SWE datasets) are widely use